Matt Welsh, an academic turned Googler, recently wrote a blog post proposing startup incubators within academia. As a once-aspiring academic turned entrepreneur, this post struck a chord with me. I agree with Matt for the most part, and wanted to voice my own thoughts.
I agree with Matt that academia is not efficient in transferring research ideas into the real world (at least from what I have seen in CS, EE, and CE). This isn’t necessarily a bad thing. The main goal of academia should be education and research (not necessarily in that order). It should not be business or profit. At the end of the day, if a university creates good students, and pushes the cutting edge in research, it is doing it’s job.
I also agree with Matt that it is worth asking if there is a role for an incubator within a university setting. My inclination is that the answer is “yes”. Here are a few reasons why I believe it may be a good idea.
1) An outlet for maximizing impact
Ask any academic about their goals, and you will hear a desire for impact. One useful way to think about impact is along a time scale. At one end of the scale is research that may have lasting impact decades from now. At the other end of the scale is research that potentially has immediate impact. There is academic research that falls at both ends of this scale, as well as everywhere in between.
If a research project has the potential for immediate impact, why shouldn’t an academic look to maximize their impact through productization? It would be good for the world. It would be good exposure for the university. It would be good for the professors and students to transfer their research into the real world.
I’m not suggesting all research be geared towards incubators, but there is a category of university research that makes sense for productization. In computer science and related fields, many research problems have the potential for immediate impact. Stanford may be one of the few universities that knows this, and is well known for spinning out influential tech companies. Luis Von Ahn has also been successful with selling ReCAPTCHA to Google, and is currently productizing some exciting research with Duolingo. There are others cases, but productizing research is not common within academia, and perhaps it should be considered more often.
2) Provides an incentive for academics besides publishing
Academia is known as a place where you publish or perish. Publications are the currency of the world. When academics talk about a desire for impact, they usually mean a desire for publications as a proxy for impact. In general, this isn’t that bad. But, there are a few negatives that come from this.
First, it incentivizes people to work on what is publishable; not necessarily on what they believe will have the biggest impact. It is easy for professors and graduate students to get caught in the system where they continually work on research that is novel, mainly for the sake of publication. How do I know? I have done it myself for many of my papers. Once you learn the system, acquire a taste for what is novel, and learn how to write, it isn’t hard to game the system. Although they probably wouldn’t publicly admit it, I know many academics also do this.
Second, it incentivizes academics to leave good ideas too early. Academics use the term LPU, or least publishable unit, to describe the minimum amount of work necessary for publication. If you know the LPU within your subfield, and want to maximize publications, it makes sense to find a hot idea, publish a LPU papers, and then leave for idea for low-hanging fruit elsewhere. Again, I know because I’ve done it. I also know others that do it. This is unfortunate because I believe in order to truly evaluate if an idea, it requires going way beyond the LPU.
It would be good for academics to have another another proxy for impact besides publication. The successful transfer of a research idea to the real world could be just this. It would provide some incentive for academics to focus research away from work that is purely publishable. It would also provide incentive to continue on research beyond publication.
3) Incubators can be profitable
This is straightforward. It is also well known that money plays a big role in academia. Universities need resources. They want to up their endowment. A university-run incubator that rolls out successful startups should also benefit monetarily.
The challenge is in running an effective incubator. Universities have an endowment to work with. They also have plenty of brains and world-experts for research. What they are missing is knowledge that incubators have learned over the years. How do you support startups at their early stages? When and how do you suggest that they pivot? When do you launch? How does do you manage your customers? This will not be easy and is not an area where universities have expertise. However, it doesn’t mean that they can’t figure it out. It is probably worth trying.
4) Growing the alumni network
High-caliber universities have high-caliber alumni networks. Graduating from the University of Illinois at Urbana-Champaign (UIUC) in Computer Engineering has given me something in common with a meaningful fraction of the engineering community. In the Bay Area, UIUC holds useful alumni networking events: I recently went to a meetup for investors and entrepreneurs. Other schools may have similar, or better, networks. For example, I’ve heard the Harvard Business School (HBS) network is a large reason for attending HBS.
Succesful incubators also have network effects. Each class of startups build a camaraderie amongst each other. Over the years, the classes build a network of businesses that help each other out. It is well-known that the growing network at Y Combinator is becoming a very powerful resource to tap into. Universities with incubators can take advantage of these network effects within their alumni network.
5) Good timing
This is also pretty straightforward. College is a great time for low-risk experimentation. College students usually do not have car payments, or a mortgage to pay. They can live off a small amount of money. It is also a great place to meet co-founders: they are everywhere!
6) Great practical education
A well-run incubator is essentially a startup bootcamp for academics. It provides students with instant education about business plans, getting funding, monetizing, dealing with customers, testing, etc. This education is valuable; especially for engineering students that know how to build, but are not exposed to the other stuff.
There is a movement towards viewing startups as experimental research. The Lean Startup Movement, a popular startup methodology, is all about reducing waste by efficiently running experiments for validated learning on your customer-product fit. Teaching students to manage these experiments would also be valuable.
In my opinion, this reason is the clincher. University is all about education. Even research and publication is about educating the world. Good universities balance theoretical education with practical/applied education. I bet that there is no better practical/applied education than bringing research into the real world. An incubator would be a great way, and perhaps the best way, for a university to make this happen.